Hi all,
while trying to provide a unit test for the PRNG in Espresso I realized that we use the ran1() algorithm from Numerical recipes, which has a rather short period. I'd like like to switch that
to ranlux[1] which has better (and proven) statistical properties and is its implementation is
part of c++11, so that we may even get rid of our own implementation in the future.
Are there any objections to this?
Regards,
Florian
p.s. Actually the original comment in src/core/random.cpp says that ran1 is bad and is only
used for seed generation, which is not true.
[1]: M. Lüscher, A portable high-quality random number generator for lattice field theory simulations, Computer Physics Communications 79 (1994) 100 (
http://arxiv.org/abs/hep-lat/9309020).
--
Florian Weik, Dipl.-Phys.,
Institut für Computerphysik, Allmandring 3, D-70569 Stuttgart
Phone: +49-711-685-67703
Public Key 0x0562F11D Fingerprint 3294 6360 EC93 37A3 BD40 F597 0BAD 3AF8 0562 F11D